Magnetic inductors and transformers are often the bulkiest, costliest, and most inefficient components in a power system. These components can limit the size and performance of a power system, and thus a larger system as a whole. In mobile, wireless, or autonomous devices, size and efficiency play an even more prominent role. For example, the power supply of an unmanned aerial vehicle (UAV) limits both mission duration and useful payload. Additionally, many consumer electronic devices demand small, efficient power supplies to enhance mobility and useful lifespan between battery recharging/replacement. Since most existing portable systems rely on dc primary power sources (such as batteries, fuel cells, and solar cells), switch-mode dc:dc power converters are commonly used to step up or down the voltage levels for various electronic subsystems. Thus, size reduction of these circuits is critical to enhance mobility and stealth.
To deliver clean, stable voltages from relatively unstable sources such as batteries or energy harvesting devices, switch-mode dc:dc power converters currently provide the highest power conversion efficiency, typically 80-95%, in the smallest form factor (dimensions). These converters employ solid-state switches (transistors and diodes) and passive energy storage elements (inductors or capacitors) to temporarily store energy from the input and then deliver the stored energy to the output at a different voltage. A standard switched-inductor boost converter is shown in FIG. 1. Referring to FIG. 1, the converter includes a transistor switch S1, a diode D1, an inductor L1, and a capacitor C1. When the transistor switch S1 is closed, magnetic energy builds in the inductor L1; when the switch S1 is opened, that energy is dumped onto the load RL. The diode D1 is used to prevent current backflow, and the capacitor C1 smoothes the ripple on the load RL. The output load voltage V0 is controlled by the percentage of time that the switch S1 is closed (i.e. the duty cycle). Standard switching speeds for commercial off-the-shelf (COTS) power converters are currently in the 100 kHz-10 MHz range.
The size and efficiency of switch-mode dc:dc converters are currently limited by the required passive components (inductors/capacitors). While the physics of semiconductor devices have enabled extreme size-reduction of solid-state devices (i.e. Moore's Law), electrical passive elements have not benefited from the same favorable scaling. Currently the silicon electronics of a dc:dc converter can be made much smaller than the required passives, which must be capable of storing a certain amount of energy. Thus, while small power IC chips are commercially available, off-chip inductors and capacitors are typically required to realize a functional power converter. For example, one related art silicon-based dc:dc buck regulator is available with a co-packaged inductor in a 3×3×1.1 mm3 package. However, the inductor dominates the package volume, and the two external capacitors used for this regulator doubles the total solution volume/mass.
To enable physically smaller passive elements, the switching frequency of the electronics can be increased. Higher switching frequencies can be used to reduce the physical size of the passive elements, since less energy must be stored per cycle to maintain the same energy transfer rate (power) at higher switching speeds. However, higher switching frequencies lead to higher power losses in both the electronics and the passive elements, and thus lower overall efficiencies. For example, with current technologies, frequencies in excess of 10-100 MHz are required to enable the passive elements to shrink to a size comparable to that of the silicon-based IC portion. At these frequencies, transistor switching losses begin to limit the overall efficiency. Additionally, frequency-dependent power losses in magnetic materials present roadblocks for realizing high-Q inductors, which again limits efficiency. Thus, existing size and performance limitations of electrical passive elements inhibit the realization of single-chip power converter solutions.
To meet the needs for fully-integrated power converter solutions, both energy density and process integrability must be addressed. Monolithic integration (fabrication of passive elements directly on the silicon) or co-packaging (interconnecting separate passive and silicon components within a package) are of equal interest, so long as high performance is achieved with small mass and volume. Accordingly, research is being conducted to achieve sufficient inductance and capacitance in a small enough form factor to enable for integration with the silicon electronics.
Researchers are actively exploring new materials and micro/nanofabrication solutions for achieving high-energy-density, low-loss inductors and capacitors. For capacitors, much attention has been focused on high-permittivity dielectrics, nanoscale dielectric gaps, and/or high-surface-area electrodes to increase the net capacitance. For example, deep trenches in silicon are being explored to increase capacitance density.
For inductors, however, there continues to be difficulties in achieving both the high inductances and high quality factors used to maintain overall converter efficiency. High permeability soft magnetic cores have attracted the most attention as means for increasing the inductance density, since the inductance ideally scales proportional with the relative permeability. While suitable magnetic films have exhibited relative permeabilities of 200-1000 in the MHz range, actual microfabricated inductors have not attained such a performance gain. This shortfall has been attributed to geometric demagnetization effects, surface defects that inhibit domain wall motion, and the difficulties in achieving fully-closed magnetic cores.
For inductors, the quality factor (the ratio of energy stored to energy lost per cycle, where Q=ωL/R) presents another major design challenge. In microscale inductors, thick electroplated coil windings are used to mitigate conduction losses, but the use of magnetic materials introduces additional power loss mechanisms, including hysteresis, eddy current, and excess loss in the magnetic cores. This additional loss is weighed against the inductance benefit offered by the magnetic core when designing the inductor. In particular, if a magnetic core introduces more energy loss than energy storage, the Q will be less than that of an air core.
The use of magnetic cores becomes more challenging as the switching frequency is increased because the magnetic eddy current losses scale quadratically with frequency. For example, one microfabricated high-inductance-density inductor using high permeability soft magnetic cores exhibited a maximum Q of only 8 at 40 MHz. For comparison, small surface-mount air-core inductors commonly used in power converter circuits typically have Q's in excess of 100. Thus, there exists a need in the art for inductor structures that exhibit high inductance density and high Q.